J Comp Physiol B (1992) 162:314-326 Joumalof " Biochemical, Compa rat lYe :~:~:';'n Physiology B =Iogy © Springer-Verlag 1992
Adaptations of the reed frog Hyperolius yiridiBayus (Amphibia: Anura: Hyperoliidae) to its arid environment
VI. The iridophores in the skin as radiation reflectors
F. Kobelt and K.E. Linsenmair* Zoologisches Institut HI, Biozentrum, Am Hubland, W-8700 Wiirzburg, FRG
Accepted November 26, 1991
Summary. Hyperolius viridiflavus possesses one complete Introduction layer ofiridophores in the stratum spongiosum of its skin at about 8 days after metamorphosis. The high reflec Members of the "super" species Hyperolius viridiflavus tance of this thin layer is almost certainly the result of (Schi0tz 1971) inhabit seasonally very hot and dry Afri multilayer interference reflection. In order to reflect a can savannas. During dry seasons, these small reed frogs mean of about 35% of the incident radiation across a aestivate unhidden mostly on dry plants. Some aspects spectrum of 300-2900 nm only 30 layers of well-arranged of their water and energy economy and of nitrogen me crystals are required, resulting in a layer 10.5 ~m thick. tabolism have been reported in Geise and Linsenmair These theoretical values are in good agreement with (1986, 1988), Schmuck and Linsenmair (1988) and the actual mean diameter of single iridophores Schmuck et al. (1988). (15.0 ± 3.0 ~m), the number of stacked platelets (40-100) As indicated by the "climate space" parameters (Gates and the measured reflectance of one complete layer of 1980), in dry seasons a skin reflectance higher than 0.7 these cells (32.2 ± 2.3 %). Iridescence colours typical of is vital during the hottest hours around noon; frogs are multilayer interference reflectors were seen after severe exposed to a heavy solar radiation load and air tem dehydration. The skin colour turned from white (0-10% peratures between 38°C and 43 °C (Kobelt and Linsen weight loss) through a copper-like iridescence (10-25% mair 1986). During acclimatization to dry season con weight loss) to green iridescence (25--42%). In dry season ditions a very conspicuous change in the wet-adapted, state, H. viridiflavus needs a much higher reflectance to yellow-brownish colour pattern of juvenile frogs can be cope with the problems of high solar radiation load seen. Skin colour is brownish-white or grey at air tem during long periods with severe dehydration stress. Dry peratures below 35 cc and becomes much paler at air adapted skin contains about 4-6 layers of iridophores. temperatures over 35-37 cc. As acclimatization pro The measured reflectance (up to 60% across the solar gresses this colour approaches brilliant white. These spectrum) of this thick layer (over 60 /-lm) is not in keep changes are accompanied by a thickening of the layer of ing with the results obtained by applying the multilayer light-reflecting iridophores (Bagnara 1976) in most parts interference theory. Light, scattered independently of of the skin from 15 /-lm up to 80 /-lm (Ko belt and Linsen wavelength from disordered crystals, superimposes on mair 1986) and a corresponding increase in the amount the multilayer-induced spectral reflectance. The initial of skin-deposited purines (Schmuck et al. 1988). parallel shift of the multi layer curves with increasing Even as long as a century ago it was assumed that light thickness and the almost constant ("white") reflectance interference was involved in the reflectance of the irido of layers exceeding 60 /-lm clearly point to a changing phores ["Interferenzzellen": Biedermann (1892); van physical basis with increasing layer thickness. Rynberk (1906)]. Denton (1970), in work on the irides cence colours of fish scales, and Land (1966, 1972) in Key words: Arid environment - Aestivation - Irido studies on the eyes of the scallop Pecten maximus, came phores - Skin reflectance - Frog, Hyperolius viridiflavus to the conclusion that iridophores in general might be taeniatus multilayer interference reflectors [see also Rohrlich and Porter (1972), Menter et al. (1979) and Frost and Robin son (1984)]. Huxley (1966) developed a method to cal culate the reflectance of these biological structures. Some morphological aspects of dermal anuran iridophores re * To whom offprint requests should be sent semble the structure of Huxley's reflectors and hence F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 315 suggest that the reflectance of the anuran skin might be The multilayer interference theory explained by this theory. Many platelets in the iridophores are arranged in The multilayer interference theory for biological systems stacks of 3-20 crystals. In recently metamorphosed wet has been developed by Huxley (1966) and Land (1966). adapted H. viridiflavus these stacks are distributed in an A detailed description of this theory is given in Appen irregular mode. In dry-adapted H. viridiflavus, however, dix B. The reflectance (R) of a multilayer interference a significant alignment is evident, many stacks being structure is given by Eqs. 9a and 9b in Appendix B. From arranged more or less parallel to the surface (Kobelt and this theory the spectral distribution of the reflected light Linsenmair 1986). However, the reflectance of multi layer can be calculated. Comparing them with the measured structures is wavelength selective, and if iridophores are spectral distribution of the light reflected from the skin pure multilayer interference reflectors their reflected light may provide clues to whether the reflecting layer has the should be coloured and not white. physical properties of a multilayer interference structure. Some investigators described three groups of irido phores in fishes (Ballowitz 1912; Connolly 1925; Becher 1929; Foster 1933, 1937; Odiorne 1933; Shanes and Nigrelli 1941; Fries 1942, 1958). One group of cells, Materials and methods leucophores, is situated in the dermis above the bony scale and reflects white light due to scattering of poorly Specimen treatment. Frogs were kept in terraria at 28/24 QC, ordered platelets. Two other groups, iridophores, one in 70/100% relative humidity (day/night) and a light regime of the dermis beneath the bony scale and another one still IlL: 13D to simulate "wet season conditions" (spraying of water and feeding daily) and "transitional conditions" (spraying and deeper and above the muscles, are iridescent due to paral feeding three times a week). "Dry season conditons" were simulated lel-ordered platelets (Denton 1970). These findings sug in an incubator at 30/20 QC, 30/70% relative humidity and IlL: 13D gest that two possible physical mechanisms are involved without feeding. Detailed descriptions of specimen treatment of in the high reflectance of light via iridophores: light H. v. taeniatus for inducing wet and dry season physiological states scattering and multilayer interference. in the laboratory have been presented elsewhere (Geise and Linsen Both physical mechanisms of reflectance provide ad mair 1988; Schmuck and Linsenmair 1988; Schmuck et aI. 1988). vantages and disadvantages: (1) Multilayer interference White reflectance measurements. For these measurements freshly structures can be manufactured very fast and at low transformed froglets were maintained under wet season conditions energy and material costs. Layers of not more than for 2 days and divided into two experimental groups. The amount 10 ~m are highly efficient light reflectors in a given spec of guanine in the skin of these froglets is less variable than in the tral range but are not very efficient across the whole solar juvenile frogs of group 3 which may have access to conditions spectrum. (2) Structures that scatter light independent of inducing dry season acclimatization. wavelength are much more effective reflectors across the Group 1: 14 froglets were kept wet-acclimatized by hydrating them every 24 h for 60 min with 500 III double-distilled water. whole solar spectrum but need much more energy and Group 2: 20 froglets were dry-acclimatized. 500 III of double material. This investigation was performed to decide distilled water was supplied for 15 min only when water losses which physical model best explains the high skin reflec exceeded 30-35% of body weight. All froglets needed a first water tance in fully dry-adapted H. viridiflavus. supply between day 10 and 12. A second supply had to be given with great individual differences mostly between day 30 and day 50 (90% of the froglets). Only about 10% received a second supply before day 30. The Kubelka-M unk-Theory Spectral reflectance measurements. For these measurements wet acclimatized juvenile H. v. taeniatus were used (age 2-4 months after Scattering of white light can be calculated according to metamorphosis). These frogs possess 1-2 complete reflecting layers the Kubelka-Munk-Theory. A summary of theories deal of dermal iridophores. Thus, errors due to iridophore-independent absorption of light deep in the body of the animal, resulting from ing with the optical characteristics of powder layers, in gaps in the reflecting layer which may be present in freshly trans which the layer is considered a continuum, has been formed froglets, are negligible. given by Kortiim (1966). (A detailed description of the Group 3: 16 juvenile frogs were maintained under transition state model and definitions of symbols are given in Appen conditions for 2 weeks and kept under dry season conditions after dix A). When S is estimated from animal data and Ro is wards. Measurement started at the beginning of the dry season measured, the thickness d of a scattering layer in the skin treatment. can be calculated from: Group 4: Ten juvenile frogs were acclimatized and kept for 2-3 months under dry season conditions before measurement. During 1 ) l-R'R this period they developed 4--6 iridophore layers, thereby matching S . d· -- R = In 0 w (4) ( Rw w Ro animals which had been collected in the field after spending a dry 1-- season of comparable duration. All frogs of groups 1-4 were mea Rw sured every 4 days to determine changes in spectral skin reflectance. They received about 50 III of double-distilled water after every The calculated data can then be compared with the mor measurement to replenish water losses which were unavoidably caused by experimental handling during measuring. phological thickness in order to see to what extent the Kubelka-Munk theory adequately describes the mea Green iridescence measurements. With increasing dehydration in all sured reflectance of the skin. groups a copper-coloured iridescence was observed, which changed 316 F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment with further dehydration to a whitish-green. As described above, reflectance. They were usually sitting motionless on their metal coloured iridescence is typical of multilayer interference reflectors plates. Thus, when the froglets were handled carefully, no an (Land 1966). This phenomenon could therefore give further clues aesthesia was required for any reflectance measurement. After to the nature of the high skin reflectance. 45 min of temperature and colour adaptation, the whitened frog Group 5: Iridescence was investigated in 13 freshly transformed together with its metal plate was transferred onto a small adjustable froglets (age and treatment identical with group 2); green irides heat conductor which was heated to a constant temperature of cence is much more pronounced in this developmental phase than 40°C. The focused beam of light was adjusted to the back of the in other age classes and usually occurred even after I week of frog and the animal placed close to the measuring hole of the sphere. dehydration. Frogs were weighed daily to determine their dehydra The skin reflectances were referred to a white standard of frog tion state. As long as their skin was white, spectral reflectance was shaped plaster cast covered with a 2-mm-thick layer of Eastman measured every 2nd day. When the copper-like colour appeared 6080 (Kodak) by a spray gun. measurements were taken daily. In contrast to the other groups, The physiological colour change is very sensitive to different froglets were given no water after measurement until green irides kinds of stress. Thus, animals that did not whiten under these cence was visible or loss of body weight exceeded 45 %. Then, after conditions were not subjected to measurements. To investigate the measurement of spectral reflectance, the frogs received 500!-l1 of effects of physiological colour change on reflectance, frogs of group double-distilled water for 60 min and their spectral reflectance was 4 were dark-adapted and measured at 20°C. measured again 2 h later. In this experiment colours were classified as follows: w: white; wc: white and spotted copper-like iridescence; Light microscopy. After reflectance measurements some froglets c: copper-like iridescence across entire dorsal skin; Cg: copper-like were precooled and double-pithed. The whole dorsal or ventral skin and spotted green iridescence; G: green iridescence across whole was fixed in 6.25% glutaraldehyde in 0.1 M monophospate buffer dorsal skin. (pH 7.2) for 30 min, then briefly washed in phosphate buffer and postfixed for I h in 1% OS04 solution (pH 7.3) (Wohlfahrt 1957). Reflectance measurements. White skin reflectance was measured The tissues were dehydrated in a graded series of acetone and using the apparatus described in Fig. la, and spectral reflectance by flat-embedded in Durcupan (Fluka) plastic resin. For light micro that shown in Fig. lb. The physiological colour change of H. v. scopy, semi-thin sections (1.5 !-lm) were made across the whole taeniatus is temperature dependent (Fig. 2a). To measure reflectance breadth of the skin and stained with methylene blue/azure II in the upper temperature range of their habitat, froglets were placed (Richardson et al. 1960). under a Perspex cover (see Fig. I) on black metal plates on a thermal The thicknesses of the reflecting layers was measured with a conductor heated to 40 QC by a thermostat. At air temperatures of calibrated ocular micrometer. The mean of 10 measurements at about 24 QC their core temperature stabilized at about 39 QC and different parts of each cross section was taken as the thickness (d) nearly all froglets became brilliant white displaying their maximum of the reflecting layer.
Monochromator
Quarz lenses
Integrating Sphere
r=oooot~'\:=:JI-Perspex Cover
b a Fig. la, b. Schematic diagram of the reflectometer: a White skin lamp. Light was measured by Si photodiodes (Hamamatsu). One reflectance was measured with a radiometer (Molectron) mounted photodiode replaced the radiometer; a second photodiode, adjusted on an integrating sphere (Earling). To simulate a more sun-like to the outlet of the monochromator, served as reference to minimize spectrum, a 30-W tungsten quartz halogen lamp (light colour 3150 errors due to fluctuations of the lamp. Spectral reflectance (Rs) was K) was run at about 10% higher voltage (Willmer and U nwin 1981). calculated from the measured pair of animal data of the two photo
The light from the lamp was focused to a small spot by quartz diodes (I, IR ) and from the pair of data for the Eastman 6080 lenses. Accuracy tested with Eastman 6080 standards was within Rs standard (Is, IRs ) to give: Rs=I·I . Accuracy was within ±5% ± I % of the measured values. b Spectral measurements were perfor IR 'Is med by mounting a 200-W mercury lamp and a monochromator of the measured values (Zeiss) on the integrating sphere (Earling) instead of the tungsten F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 317
0.60
* T * liT Q) 0.55 u c 1/- (\l U Q) ~ .} 1 c :;;; Cl) * l 0.50 l I~I Fig. 2. a Correlation between skin re· - flectance and core temperature. Frogs T/1 of all age classes exhibit a correlation l""T!--~ group 4 .1. --1 between core temperature and skin col· 0.45 our. This is pronounced in dry-adapted 10 15 20 25 30 35 40 45 frogs of group 4 which try to a Temperature [QC] match the black colour of their seats between 20 and 30 QC. Note the signifi· cant rise of skin reflectance above 30 QC. b Increase of skin reflectance under prolonged dry season conditions. Skin reflectance increased in a logarith· 0.5 * mic pattern. Note the plateau between 15 day 8 and day 12 in group 2. The for R2 = 0.061 ·In t + 0.254 mulas represent the relationships of the Q) u r = 0.962; P 40.------~ * 30 E ::J. (/) (/) Q) Fig. 3. Thicknesses of the reflecting lay· C ..l<: er calculated from the Kubelka·Munk U 20 theory and measured from skin sec £... tions. All measured thicknesses of the Q) reflecting layer are below the calculated >- (\l . 0····· 0 group 1 calculated curves. Differences are especially high ..J at the beginning of the experiment. fiJ A-A group 2 calculated • * Significant difference between mea 10 T r • group 1 measured * sured and calculated thicknesses (t-test, A group 2 measured measured mean against the correspond· I ing theoretical value, P < 0.05) 0 10 20 30 40 50 60 70 Time 318 F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment Results All thicknesses derived from light microscopy are smaller than the values calculated from Eq. 4. Differences White reflectance are especially high at day 4 (70.6%) and day 8 (43.5%). The mean differences (± standard deviation) are (Fig. 3): At air temperatures between 20 and 30 °C, the froglets Group 1: 21.3±17.3%,n=4; show a colour matching that of the background. This Group 2: 12.0± 8.0%, n= 7. colour changes to a brilliant white if the air temperature exceeds 35°C (Kobelt and Linsenmair 1986). The cor Spectral reflectance relation of skin colour and core temperature is given in Fig. 2a. A detailed description of these results will be The spectral reflectance of group 3 has a broad maximum presented elsewhere (F. Kobelt and K.E. Linsenmair, in between 800 and 1100 nm, and two smaller maxima at preparation). 360 nm and 520 nm. It increases steeply below 340 nm. Usually the iridophores are shaped like globes with a After 20 days under dry season conditions skin reflec diameter of 15 ± 3 Ilm. When the number of iridophore tance is shifted about 0.12, but the shape of the curve is layers exceeds three, these cells become elliptically shaped almost identical. The smaller maxima remain at the ini and the diameter is highly variable (Ko belt and Linsen tially measured wavelengths (Fig. 4a). mair 1986). Skin reflectance data of groups 1 and 2 were Group 4 has a spectral reflectance which is almost presented in detail by Schmuck et al. (1988). A brief constant across the whole measured spectrum. A steep summary of these data is given in Fig. 2b as an aid to absorption band is seen at 320 nm. From the differences understanding the following results. in spectral reflectance between white- and brownish- 0.7 0.6 ~ 0.5 c ell 0 Q) 0.4 ~ '- c :.;;:: (f) 0.3 0.2 • .. ···-·day 11 (n=16) .-. day 20 ~ 0.1 a Fig. 4. a Spectral skin re· flectance of group 3 frogs 0.8 0.8 before and after 20 days under dry season con· 0.7 0.7 ditions. The shape of the curve did not change sig. nificantly; it is shifted 0.6 0.6 about 0.12 almost parallel Q) c to the curve for day 1, but 0 0 c 0.5 0.5 the IR-maximum becomes ~ 'e- distinctly broader. b Spec· 0 0 Q) Cl) tral skin reflectance and 0.4 0.4 ..0 ~ ell absorption of group 4 c c frogs at 20 °c and 40°C. :.;;:: 0.3 0.3 :.;;:: (f) Frogs with a reflecting lay- (f) er of a thickness exceeding 0.2 0.2 60 Ilm had a high reflec· tance across the whole •. spectrum; melanin absorp . 0.1 ...... 0.1 tion was calculated by sub- • ' , ...... tracting the curve at 40°C 0.0 0.0 from that at 20 °c 200 400 600 800 1000 1200 b Wavelength [nm] F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 319 Fig. Sa, b. Skin colour of a dry-adapted juvenile and b adult taken in Comoe National Park. Ivory Coast. b Skin colouration of H. viridiflavus. a Juvenile frogs aestivating in a typical position and adult frogs is entirely different from that of juvenile frogs, with a exhibiting light-adapted skin colour at air temperatures above 35 QC variable black and yellow-whitish pattern and red inner sides of (al) and dark-adapted skin colour below 35 QC (a2). Pictures were limbs and toes 120 y 120.9-25.9·t+4.2·t2 -O.23·t3 5t = 1 H2 0- r = 0.986; P < 0.01 supply 110 rehyd",.d 15 ! i ..·::1 .: 100 .:: ,w . . " ~ ...... ?f2. 90 .. . E Ol 80 .... 'wc (J.) ..... ~""c'" ~ Fig. 6. Body weight changes of group 5 >- ...... ~ ... ~~.. -a 70 frogs during dehydration. Open sym 0 . . bols represent body weight of frogs [(} ~ ...... which died before the next measure 60 ment. The colour classification in rela tion to dehydration state is depicted on 50 0 0 0 the right side. The body weight of five 0 • ,"~i"o" f G o dead frogs specimens 2 h after rehydration is also 0 40 shown 0 2 3 4 5 6 7 8 9 10 11 Time [d] coloured animals the absorption due to the physiological whitish pattern and red inner sides oflimbs and toes (Fig. colour change can be calculated. Dark skin absorbs 5b). Their skin reflectance also increases with rising tem mainly light of the visible part of the spectrum, IR-radia perature, but the reflectance range (0.38-0.53) is signifi tion only between 0.08 and 0.1 (Fig. 4b). These laborato cantly lower than in dry-adapted juvenile frogs. ry frogs exhibit skin colouration and skin reflectance comparable to that of field frogs of the same age (2-4 Green iridescence months after metamorphosis; Fig. 5a). Skin colouration of adult frogs is entirely different Iridescence started mainly at the lower hind legs of group from juvenile frogs with a variable black and yellow- 5 froglets and progressed with increasing dehydration to 320 F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment . 0.5 '.~...... ·····-.white (n=13) 4~4:: .. ! ········copper (n=11) ... . .-. green (n= 7) ,, . , . : .... ~. 0.2+-~-r--~~~'-~-r __~~~~~-r __~~~~~-r __~ a 0.5 ..... - white ...... green .-. rehydrated (n=5) Q) t) 0.4 c: Fig. 7a, b. Changes of skin reflectance co t5 of group 5 frogs: a during dehydration Q) and b after rehydration. With increas ~ ing dehydration the broad maximum c: between 750 nm and 850 nm diminishes ". ~ : and the green maximum at 520 nm be· Cl) 0.3 ..... ' .. comes more pronounced (a); 2 h after rehydration the pronounced green maximum at 520 mm was lowered (b) and the spectrum approximately re· turned to the copper colour depicted in (a) 0.2+-~-r--~.-~.-~-r--~.-~.-~-r--~.-~.-~-r--~ 200 400 600 800 1000 1200 b Wavelength [nm] the back of the frog. The change of the iridescence colour visible spectrum are much lower and less pronounced. A is closely related to weight loss due to evaporation and new maximum at 620 nm is found. The broad IR maxi energy depletion (Fig. 6). After a weight loss of about mum around 800 nm has almost diminished. With in 15%, copper-coloured iridescence is at first weak but is creasing dehydration, the UV maximum is shifted further clearly visible when weight loss reaches 22%. A weak to the UV and the small maximum at 520 nm becomes whitish-green iridescence is seen after a weight loss of clearly pronounced. The broad maximum around about 31 %. Frogs with a weight loss of more than 40%, 800 nm has diminished and the two IR maxima at 980 which is close to the lethal limit of this age class, have a and 1080 nm are lowered (Fig. 7a). The reflectance of clear whitish-green iridescence. At the end of the experi green iridescing froglets recovered to a copper or copper ment more and more frogs come close to their individual whitish skin colour within 2 h after rehydration (Fig. 7b). dehydration limit, and despite rehydration three of the Survivors showed the typical white skin colour the day five whitish-green iridescent frogs with a weight loss of following rehydration. more than 40% died within a week after rehydration (Fig. 6). In the spectrum of the whitish froglets a high maxi Discussion mum in the UV at 320 nm is seen. Smaller maxima are found at 420, 460, 520 and 560 nm followed by a broad In freshly transformed froglets purines are synthesized maximum between 750 and 850 nm and two smaller regardless of the environmental conditions. Froglets ex maxima at 980 and 1080 nm. In the spectrum reflected by posed to dry season conditions almost doubled the rate the copper-coloured frogs the UV maximum is shifted to of purine synthesis after a 10% loss of the initial body the UV range of the spectrum and the maxima of the weight during the first 4 days of the transition state F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 321 (Schmuck and Linsenmair 1988). The plateau between Spectral reflectance. Many anurans have one complete day 8 and day 12 coincides with a decrease of the respira layer of iridophores in the stratum spongiosum. This tory quotient (Geise and Linsenmair, in prep.). These layer, together with melano- and lipophores, is the sub phenomena seem to mark the point where the physiology strate of physiological colour change (Fox 1953; Cott of the frog changes from transition state to dry-adapted 1957; Fox and Vevers 1960; Nopp 1964; Fingerman state, which is further discussed in Schmuck et al. (1988). 1965; Bagnara 1966, 1976; Bagnara and Hadley 1973; Three different reasons for the tremendous build-up Bagnara et al. 1968, 1978; Nielsen 1978). The high reflec of purines in the skin of some Hylidae and Hyperoliidae tance of the thin reflecting layer in the skin of recently during aestivation periods are discussed in the literature: metamorphosed H. viridiflavus (groups 1 and 2) is almost certainly the result of multilayer interference reflection (1) Skin permeability. Some authors speculate from their (Frost and Robinson 1984; Kobelt and Linsenmair 1986). results that an unequal distribution of iridophores in The mean thickness of one iridophore is 15.0 ± 3.0 /lm dorsal and ventral skin corresponds to differences in the (n = 20 frogs), the measured reflectance of one complete evaporative water loss of these types of skin (Drewes et reflecting layer is 0.322 ± 0.023 (n = 20), and the mea al. 1977; Yorio and Bentley 1977; Withers et al. 1982). sured spectral maxima are in the range 800-1100 nm (see This conclusion is in agreement with the results of Kutchai Fig. 4). Calculation of comparable data from our model and Steen (1971) who found that the permeability of the (Eqs. 9-16) revealed that in order to reflect a mean of swim bladder of the conger eel (Conger conger) is only about 0.35 across a spectrum of 300-2900 nm at Ao = 817 10% that of connective tissue. The silvery layer produc only 30 layers of well-arranged crystals are needed, re ing this effect contains 13% guanine (dry tissue). sulting in a reflecting layer thickness of 10.5 /lm. These values are in good agreement with the measured data. The small differences between calculated and measured (2) Nitrogen metabolism. Some anurans aestivating values are most probably caused by discontinuities in the above ground use purines to redress osmotic problems reflecting layer (e.g. morphological structures such as resulting from urea accumulation in the body fluids by gland ducts), leading to absorption of light in the irido internal deposition of nitrogenous waste products with ph ore-free gaps in the skin. Spectral curves with shapes out fluid loss (Shoemaker et al. 1972; Balinsky et al. closely comparable to that of H. viridiflavus are found 1976; Schmuck and Linsenmair 1988; Schmuck et al. elsewhere in the literature (Tracy 1972; Carey 1978; 1988). Gates 1980), but the reflectance is much lower because a much greater amount of light is absorbed by melano (3) Optical properties. Besides physiological colour phores in the investigated species (Rana pipiens, Bufo change (Bagnara 1976), purines seem to be used in boreas). anurans aestivating above ground for thermoregulation From the multilayer interference theory spectral by protecting the animals against the harmful solar curves can be calculated that conform in some charac radiation load (Kobelt and Linsenmair 1986; F. Kobelt teristic traits with the curves derived by averaging the and K.E. Linsenmair, in preparation). This paper measured spectral data of group 3 (Fig. 8). One charac proposes a possible physical basis for the high skin reflec teristic of the theoretical curves is that the addition of tance, and this is discussed in detail in the following crystal layers gradually influences the shape of the sections. theoretical curves only in the range 1-30 layers. Above 0.8 ~ 0.12 ...l ... 0.7 30 layers the influence of a new layer is almost negligible. stances significantly decreases with increasing wavelengths The same characteristic holds true for the measured from about 700 nm to 2500 nm (Gates 1980). This de curves; their spectral shape remains unchanged when the crease might be the main reason for the observed dif reflecting layer is increased from 15 Ilm to 45 Ilm. How ferences above 700 nm because it is accompanied by a ever, the fraction of "white" light scattered independent displacement of "-0 to smaller wavelengths, resulting in a ly of wavelength by the thickened layer increases and at lower reflectance at the calculated wavelength. Unfor a given diameter adds a constant fraction of light to the tunately, no refractive index data above 700 nm for pu multilayer-induced reflectance. As a result the reflectance rines were available. Therefore, this error could not be is shifted up with only very small changes of its spectral corrected in the calculations. composition (Fig. 4a). Such a shift is seen in the spectral In electron microscope studies Kobelt and Linsenmair reflectance at day 1 and day 20 in group 3. Both curves (1986) found a highly ordered arrangement of crystals have almost the same shape and differ by about 0.12 from parallel to the skin surface in dry-adapted frogs. The each other (Fig. 8). This is comparable with the 0.15 calculations were made under the assumption of parallel increase of the white reflectance of group 2 during the layers of crystals, and the relatively good fit of theoretical same period in our dry season conditions (Fig. 2). and measured curves indicates that this assumption is not The measured and the calculated curves show similar seriously at fault. Since the reflectance curves of day 1 characteristics, i.e. both rise in the UV and have a broad can also be calculated from the multilayer theory, wet maximum in the IR. The differences might be brought adapted frogs may also temporarily exhibit a high paral about by absorption in the skin. Besides xanthophore lel orientation of crystals in their iridophores. Nielsen pigments, many other organic substances strongly ab (1978), working in light-adapted skin of Hyla arborea, sorb light in UV below 320 nm [absorption maxima of reported a parallel orientation of the crystals to the skin purines are between 240 and 290 nm (Stahl 1967)] and surface, whereas in dark-adapted skin the crystals were the differences in the UV part of the curves can possibly mainly randomly distributed. He found iridophores with be explained by absorption. Whether absorption is also more or less parallel crystals predominantly in connec responsible for the differences in the IR part of the spec tion with light-coloured backgrounds. The orientation of trum remains unclear. Possible alternative factors might the crystals is presumably performed by a network of be: (1) Water absorbs radiation weakly above 700 nm filaments between the crystals (Rohrlich and Porter and strongly above 1100 nm. However, a water layer of 1972). These results give rise to the assumption that 60 j.lm thickness (thickness of the iridophore layer plus changing the arrangement of the crystals is a fast process epidermis in group 3) alone cannot cause these differ allowing quick responses to short-term changes in en ences. (2) Porter (1967) found that a shed skin of Uta vironmental conditions. Very likely the iridophores of stansburiana about 20 j.lm thick absorbed about 0.27 of H. viridiflavus also have this ability. This problem is the incident radiation between 700 nm and 2600 nm, currently under investigation. which was mainly a result of absorption by keratin. A system of highly parallel oriented crystals should However, a cell layer of comparable keratinization in the reflect coloured light, but the reflected light of the irido skin of H. v. taeniatus, the stratum corneum, is only phore layer in dry-season frogs is yellow-white or bril about 2 j.lm thick and its IR absorption amounts to not liant white, indicating a broad-band spectral reflectance. more than 0.012. (3) The refractive index of many sub- These differences of the spectral reflectance might be 0.8 -de =201 nm . - - de = 150 nm 0.7 ..... de = 100 nm ..... de = 50 nm \ Q) \ 0 c 0.6 \ ('\l \ 0 \ Q) :;::: \ \ ~ 0.5 \ Fig. 9. Theoretical displacement of 1.,0' Q) , \ > , , With decreasing de the IR maximum is '"§ \ \ lowered and displaced more and more ID 0.4 \ to the visible spectrum. The displaced er: , , \ , maximum shifts into the visible spectral ...... " range after a more than 40% decrease 0.3 ...... of de. The smaller maximum at 350 nm ...... :.,: is also shifted into the UV region. All curves were integrated in the range of 0.2+-~.-~-.~-.--~r-~'-~-r~-.~-.r-~.-~.-~~ de ±0.25· de 200 400 600 800 1000 1200 Wavelength [nm] F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 323 explained by irregularities of the parallel arrangement of hypothesis is supported by Land (1966), who found a the crystals in the iridophores. Usually dry-season frogs shift from the initial blue-green to blue after suspending show a good orientation of crystals in light-adapted the eye of Pecten maximus in hyperosmotic sucrose, and iridophores; however, not all crystals are arranged in a shift through yellow to orange in distilled water. These stacks and not all stacks are parallel to the skin surface colour shifts were reversible after replacement of the eye (Kobelt and Linsenmair 1986). These discontinuities en in seawater. Such a displacement can be calculated from hance scattering of the incident light within the reflecting the multilayer theory, but the measured curves (Fig. 7) layer, and the light becomes more diffuse. In addition, do not agree very well with the calculations (Fig. 9). the cytoplasm interspaces are very variable, leading to more than one Ao, and the properties of the layer come White reflectance. In H. viridiflavus a skin reflectance of close to the interference mirrors of Heavens (1960). This more than 0.7 is needed to meet the problems of high can be seen by examining the skin under a microscope. solar radiation load in the field (F. Kobelt and K.E. In light-adapted skin tiny spots of red, yellow or green Linsenmair, in preparation). The reflectance of thick iridescence of similar diameter to that of a dorsal irido iridophore layers (Fig. 4b) does not conform to the re phore can be seen. These iridescent spots are probably sults obtained by applying the multilayer interference the result of multilayer interference within iridophores theory (Fig. 8). Light which is scattered independent of with highly ordered crystal systems at the top of the wavelength from disordered crystals superimposes on reflecting layer. The different colours are probably the multilayer-induced spectral reflectance. The initial caused by the different angles of incidence and by dif parallel shift of the multilayer curves with increasing ferences in the cytoplasm interspace thickness. As a re thickness (Fig. 8) and the almost constant ("white") sult, "white" light may also be reflected from a multilayer reflectance of a layer thicker than 60 )lm (Fig. 4b) clearly when the incident light becomes diffuse and the crystals points to changing physical properties with increasing are not perfectly spaced quarter-wavelength reflectors layer thickness. (Heavens 1960). Below a thickness of 15 )lm the difference between the Under certain conditions the reflected light of the measured and the calculated thickness from the whole skin of H. viridiflavus is iridescent in colours typi Kubelka-Munk theory averages 21.3%. Just after meta cal of multilayer interference reflectors. Iridescence col morphosis the froglets possess an incomplete iridophore ours are seen in the skin of dehydrated frogs after more layer covering 50-80 % of the dorsal stratum spongiosum than about 22 % weight loss. A result of dehydration (Ko belt and Linsenmair 1986). Through the uncovered might be a shrinking of the cytoplasm, leading to reduced areas a considerable amount of the incident radiation can cell volume and decreased mean thickness of the cyto penetrate deep into the body where it is absorbed. Ac plasm interspaces. This probably leads to a parallel cording to the Kubelka-Munk theory absorption of the orientation of the crystals within the shrunken cell result reflecting material is not taken into consideration. ing in a highly ordered structure with even better mul Therefore, skin reflectance at this stage is much lower tilayer interference properties than the original cell. than calculated from this theory and may differ from the Additionally, the reduced thickness of the cytoplasm calculated values by up to 70.6% at day 4 and 43.5% at interspaces seems to be responsible for the overall dis day 8 (Fig. 3). In group 1 and group 2 after about day placement of Ao from the IR to the visible spectrum. This 12 the reflecting layer is complete and its thickness is 0.7 0.6 0.5 Q) ...... 0 .. " C ...... 0.7) for aestivation, The Multi/ayer Interference Theory costing much more in material and energy than a mul tilayer structure, is not needed before the periods of Light reflected from a single platelet is coloured because complete dryness exceed about 5-7 days. In the tran of constructive interference. The wavelength of the col sitional period the froglets are able to get enough energy our which is best reflected (/co) is given by: by feeding and to replenish water lost by evaporation or 4 /co = . d· Vn2 - sin2a (5) used for evaporative cooling. Usually the froglets have 2· z-1 sufficient time, material and energy to establish the re quired high skin reflectance in this period. where z = 1, 2, 3, ... m. F. Kobelt and K.E. Linsenmair: Adaptations of H. viridiflavus to its arid environment 325 Taking the first order (z= 1) and for light incident nor 21t cD = - . n . d . cos e (15) mally on the platelet (a = 0) one obtains for the optical k A K K thickness (Ll) of the platelet imbedded in cytoplasm n sin a [n = refractive index, d= platelet thickness, K= crystal, k (16) e = cytoplasm; Land (1966)]: (6) A single quarter-wavelength crystal reflects between 6% and 9% of the incident light over the whole visible spectrum. However, in a multilayer system the variation A multilayer interference reflector is manufactured by of reflectance with wavelength becomes more pronounced using films of alternating high and low refractive index. with increasing number of layers, the bandwidth be If a system of crystals is to function in the manner of a comes narrower and more sharply defined, and the ini multilayer interference reflector, both the crystals and the tially unsaturated colour becomes correspondingly more spaces between them should have optical thicknesses of intense. With 30 ideally structured layers the reflectance Ao/4. Huxley (1966) has derived a method to calculate the approaches 100% around Ao. The wavelength depen reflectance of quarter-wavelength systems. The reflec dency can be reduced by varying the thicknesses of the tance of one platelet is given by Fresnel's formula: constituent layers. Interference reflectors with about 95 % reflectance over the whole visible spectrum have been (7) manufactured by applying this technique (Heavens 1960). In a biological system of platelets the incident light is where r = reflectance of one platelet, 1= reflected light, scattered in the epidermis and at the sides of the crystals 10 = incident light and e = angle of I. For light incident which are not ideally parallel to each other. Therefore, normally on the platelet (a = 0) : the light within the multilayer system can be considered as being diffuse. The crystal thickness (dK) ranged from 0.132 to 0.1821-1-m and the cytoplasm interspaces (de) r-_(1-n)2 -- (8) l+n from 0.153 to 0.253).lm (Kobelt and Linsenmair 1986), resulting in variations of the optical thickness of the The reflectance (R) of the quarter-wavelength system crystals (nKdK) and the intermittent cytoplasm (ncde) with p crystal layers for real and complex numbers of Il around the ideal optical thickness (nodo) for both sys (see Eq. 13) is: tems. This was taken into consideration by numerical integration of Eq. 9 over a, dK and de with a computer (Fig. 8). Ao was calculated to be 817 nm from Eq. 6 with a) f1 real R = 2 (9a) 1-p nc= 1.33 (Land 1968), nK = 1.81 and dK =0.150).lm 1+4m2 .-~ (l-m)2 (Ko belt and Linsenmair 1986). 1 Acknowledgements. This study was supported by the Deutsche For b) f1 complex R = --.,------(9b) schungsgemeinschaft, Research Grant Li 159/11-1/2. 1 + (P2 - 1) . cosec2 (2p lebenden Chromatophoren der Fischhaut in auffallendem Licht. Menter DG, Obika M, Tchen TT, Taylor D (1979) Leucophores Z Wiss Mikrosk 46: 89-124 and iridopheres of Fundulus heteroclitus: biophysical and ultra Biedermann W (1892) Uber den Farbwechsel der Frosche. Arch structural properties. J Morphol 160: 103-120 Ges Physiol 51 :455-508 Nielsen HI (1978) Ultrastructural changes in the dermal chromato Bone Q, Denton EJ (1971) The osmotic effects of electron micro phore unit of Hyla arborea during color change. Cell Tissue Res scope fixatives. J Cell Bioi 49: 571-581 194:405-418 Carey C (1978) Factors affecting body temperature of toads. Nielsen HI, Dyck J (1978) Adaptation of the tree frog, Hyla cinerea, Oecologia 35: 197-219 and the role of the three chromatophore types. J Exp Zool Connolly CJ (1925) Adaptive changes and colour of Fundulus. Bioi 205:79-94 Bull 48: 56-77 Nopp H (1964) Melanine und ihre Rolle im tierischen Organismus. Cott HB (1957) Adaptive colouration in animals. Methuen, London Verh Zool Bot Ges Wien 103/104: 16-54 Denton E (1970) Review lecture: on the organisation of reflecting Odiome JM (1933) The occurrence of guanophores in Fundulus. surfaces in some marine animals. Philos Trans R Soc London Proc Natl Acad Sci USA 19:535-540 Ser B: 258: 285-313 Porter WP (1967) Solar radiation through the living body walls of Drewes RC, Hillmann SS, Putnam RW, Sokol OM (1977) Water, vertebrates with emphasis on desert reptiles. Ecol Monogr nitrogen and ion balance in the African tree frog Chiromantis 37:273-296 petersi Boulanger (Anura: Rhacophoridae) with comments on Romeis B (1968) Mikroskopische Technik. Oldenburg Munchen, the structure of the integument. J Comp Physiol 116: 257-267 Wien Fingerman M (1965) Chromatophores. Physiol Rev 45: 296-339 Rohrlich ST, Porter KR (1972) Fine structural observations relat Fischer U (1980) Warmereflektierendes Plexiglas. R6hm Spektrum ing the production of co10r by the iridophores of a lizard, Anolis 21: 33-34 caroliniensis, J Cell BioI 53: 38-52 Foster KW (1933) Colour changes in Fundulus with reference to Schiotz A (1967) The treefrogs (Rhacophoridae) of West Africa. colour changes of iridosomes. Proc Natl Acad Sci USA Spolia Zool Mus Hauniensis 25: 222-229 19:535-540 Schiotz A (1971) The superspecies Hyperolius viridiflavus (Anura). Foster KW (1937) The blue phase in the colour change of fish with Vidensk Medd Dan Naturhist Foren Khobenhavn 134:21-76 special reference to the role of guanine deposits in the skin of Shanes AM, Nigrelli RF (1941) The chromatophores of Fundulus Fundulus heteroclitus. J Exp Zool 77: 169-213 heteroclitus in polarized light. Zoologica (N Y Zool Soc) Fries EFB (1942) White pigment effectors (leucophores) in killi 26:237-243 fishes. Proc Natl Acad Sci USA 28: 396-401 ~chmuck R, Linsenmair KE (1988) Adaptations of the reed frog Fries EFB (1958) Iridescent white reflecting chromatophores . Hyperolius viridiflavus (Amphibia: Anura: Hyperoliidae) to its (antaugonophores, irido1eucophores) in certain te1eost fishes, arid environment: Ill. Aspects of nitrogen metabolism and particularly in Bathygobius. J Morphol 103: 203-254 osmoregulation in the reed frog, Hyperolius viridiflavus taeniatus, Fox DL (1953) Animal biochromes and structural colours. Cam with special reference to the role of iridophores. Oecologia bridge University Press, Cambridge \~ 75:354-361 Fox HM, Vevers HG (1960) The nature of animal colours. Oxford t:Schmuck R, Kobelt F, Linsenmair KE (1988) Adaptations of the University Press, Oxford .' reed frog Hyperolius viridiflavus (Amphibia: Anura: Hype Frost SK, Robinson SJ (1984) Pigment cell differentiation in the roliidae) to its arid environment: V. Iridophores and nitrogen fire-bellied toad, Bombina orientalis: I. Structural, chemical, and metabolism. J Comp Physiol B 158: 537-546 physical aspects of the adult pigment pattern. J Morphol Shoemaker VH, Balding D, Ruibal R, McClanahan LL (1972) 179:229-242 Uricotelism and low evaporative water loss in a South American Gates DM (1980) Biophysical ecology. Springer, New York frog. Science 175: 1018-1020 Heidelberg Berlin Stahl E (1967) Dunnschicht-Chromatographie. Springer, Berlin Geise W (1987) Leben unter Extrembedingungen: Untersuchungen Heidelberg New York zur Aestivationsphysiologie und zur Variabilitat im Lebens Tracy CR (1972) A model of water and energy dynamic inter zyklus beim afrikanischen Riedfrosch, Hyperolius viridiflavus relationships between and amphibian and its environment. (Anura: Hyperoliidae). Dissertation, Faculty cif Biology Uni PhD thesis, University of Wisconsin versity of Wurzburg Tracy CR (1975) Water and energy relations of terrestrial am Geise W, Linsenmair KE (1986) Adaptations of the reed frog phibians: in sights from mechanistic modeling. In: Gates DM, Hyperolius viridiflavus (Amphibia: Anura: Hypero1iidae) to its Schmerl R (eds) Perspectives in biophysical ecology. Springer, arid environment: Il. Some aspects of water economy of New York Berlin Heidelberg, pp 325-346 Hyperolius viridiflavus nitidulus under wet and dry season con Tracy CR (1976) A model of the dynamic exchanges of water and ditions. Oecologia 68: 542-548 energy between a terrestrial amphibian and its environment. Heavens OS (1960) Optical properties of thin films. Rep Prog Phys Ecol Monogr 46:293-326 23: 1-65 Van Rynberk G (1906) Uber den durch Chromatophoren bedingten Huxley AF (1966) A theoretical treatment of the reflexion of light Farbwechsel der Tiere (sog. chromatische Hautfunktion). Ergeb by multilayer structures. J Exp Bioi 48: 227-245 Physiol BioI Chem Exp Pharmakol 5: 347-571 Kobelt F, Linsenmair KE (1986) Adaptations of the reed frog Willmer PG, Unwin DM (1981) Field analyses of insect heat bud Hyperolius viridiflavus (Amphibia: Anura: Hyperoliidae) to its gets: reflectance, size and heat rates. Oecologia 50: 250-255 arid environment: 1. The skin of Hyperolius viridiflavus nitidulus Withers PC, Hillmann SS, Drewes RC, Sokol OM (1982) Water loss in wet and dry season conditions. Oecologia 68: 533-541 and nitrogen excretion in sharp-nosed frogs (Hyperolius nasutus: Kortum G (1969) Reflexionsspektroskopie, Springer, Berlin Anura: Hyperoliidae). J Exp BioI 97: 335-343 Heidelberg New York Wohlfarth KE (1957) Die Kontrastierung tierischer Zellen und Kutchai H, Steen JB (1971) The permeability of the swimbladder. Gewebe im Rahmen ihrer elektronenmikroskopischen Unter Comp Biochem Physiol 29: 119-123 suchung an ultra-dunnen Schnitten. Naturwissenschaften Land MF (1966) A multilayer interference reflector in the eye of the 44:287-288 scallop Pecten maximus. J Exp Bioi 45:433-477 Yorio T, Bentley PJ (1977) Asymmetrical permeability of the in Land MF (1972) The physics and biology of animal reflectors. Prog tegument of tree frogs (Hylidae). J Exp Bioi 67: 197-204 Biophys Mol Bioi 24:75-106